List your company
ResearchFormerly WizWand

Language Understanding on C-Eval

87.7C-Eval ScoreQwen2-57B-A14B
24.280840.745457.2173.6746Aug 24, 2023Jan 18, 2024Jun 14, 2024Nov 9, 2024Apr 6, 2025Sep 1, 2025Jan 27, 2026
Updated 6mo ago

Evaluation results

MethodLinks
Qwen2-57B-A14BArchitecture=MoE, # Act Params=14B, # Params=57B2024.0787.7
Qwen1.5-32BArchitecture=Dense, # Act Params=34B, # Params=34B2024.0783.5
Qwen2-7BParameters=7B2024.0783.2
Qwen1.5-7BParameters=7B2024.0774.1
KEELArchitecture=512 Layers / 3B Params, Peak Learning Rate=4.5 x 10^-3, Pre-training Tokens=1T, Evaluation Protocol (Shots)=5-shot2026.0169.5
CPTSize=70B, Evaluation Mode=5-shot2024.0968.36
Llama-3Size=70B, Evaluation Mode=5-shot2024.0967.75
Pre-LNArchitecture=512 Layers / 3B Params, Peak Learning Rate=3.0 x 10^-3, Pre-training Tokens=1T, Evaluation Protocol (Shots)=5-shot2026.0166.2
Qwen-7BModel Size=7B, Status=final released2023.0863.5
Llama-3-SynEevaluation_mode=Few-shot2024.0758.24
Baichuan2-7BModel Size=7B2023.0854
InternLM-7BModel Size=7B2023.0852.8
ChatGLM2-6BModel Size=6B2023.0851.7
CPTSize=8B, Evaluation Mode=5-shot2024.0951.12
Qwen-VLInitialization=Intermediate Qwen-7B checkpoint2023.0851.1
Llama-3-ChineseSize=8B, Evaluation Mode=5-shot2024.0950.5
Llama-3-Chinese-8Bevaluation_mode=Few-shot2024.0750.14
Llama-3-8BParameters=8B2024.0749.5
Llama-3-8Bevaluation_mode=Few-shot2024.0749.43
Llama-3Size=8B, Evaluation Mode=5-shot2024.0949.38
Qwen-7BModel Size=7B, Status=intermediate, Usage=LLM initialization for Qwen-VL2023.0848.5
Mistral-7BParameters=7B2024.0747.4
MAmmoTH2-8Bevaluation_mode=Few-shot2024.0746.56
Gemma-7BParameters=7B2024.0743.6
Baichuan-7BModel Size=7B2023.0842.8
Mistral-7B-v0.3evaluation_mode=Few-shot2024.0742.74
DCLM-7Bevaluation_mode=Few-shot2024.0741.24
LLAMA2-7BModel Size=7B2023.0832.5
Galactica-6.7Bevaluation_mode=Few-shot2024.0726.72